The ubiquity of smart phones and tablet computers, as well as text messages, chat functions and other non-traditional means of digital communication, present trade-surveillance managers with the technological challenge of how to monitor activity conducted via such ‘other’ mechanisms.

For the biggest global banks that need to police their ecosystem for bad behavior, and may be looking to leverage the business intelligence gleaned from that mandate to generate revenue, it’s a work in progress.

“The industry as a whole is grappling over how best to use this data,” said Theo Hildyard, head of solution marketing at Software AG. “The first problem you have to solve when dealing with these other data sources is how to integrate them, as in physically build the data and plug in the APIs (Application Programming Interfaces) that can suck the data in. Then you have to have an architecture that’s scalable enough to actually make use of the data.”

The next generation of surveillance systems — which is being developed and refined today — will be capable of analyzing in real time the ‘fire hose’ of market and trade data and, increasingly, social media, emails, instant messages, news headlines, audio data from phone calls, or video data from surveillance activities.

“The ability to integrate unstructured data such as chat rooms, emails, IMs and phone including mobile phone data into the mix of market data, and orders and trades data to more accurately pick out patterns of potential collusion and front running/insider trading is something that the industry is actively researching right now,” said Magnus Almqvist, senior product specialist in SunGard’s capital markets business.

To monitor and make sense of reams of fast-moving data, firms must connect to multiple, disparate, live data sources, to ingest and digest all of these data streams, and to handle the peaks and troughs of the velocity, all the while detecting actionable patterns. The ability to monitor activity across multiple platforms requires synthesizing various data elements to create a comprehensive, holistic view across the entire organization.

While tasked with aggregating and understanding data from disparate sources, surveillance technology also must continue to pull everything forward. Knowing about a trade X hours after it happened may have been acceptable for yesterday’s systems, but today’s standard is X minus Y hours, and tomorrow’s will be X minus Y minus Z. This applies to trade surveillance across all financial ecosystems.

“The idea is if we see something happening that is disruptive to the market, we can either interrupt that or at least detect that much sooner than most of the surveillances that are currently run today,” said Mary Brienza, chief executive officer of NYSE Regulation Inc., a subsidiary of IntercontinentalExchange’s NYSE. “The object is to move everything up in time.”

NYSE Regulation will directly perform market surveillance, investigation and enforcement functions for NYSE Group’s three equities exchanges and two options exchanges upon the expiration of its contract with the Financial Industry Regulation Authority at the end of the year.

Once a pattern or suspicious behavior has been identified, the next step is to look at communication through every available channel to find evidence of intent and/or collusion.

For example, say a firm’s prop desk is trading against its investment advice, or trading against its clients. “To detect this efficiently and systematically requires a holistic view on investment advice, client transactions and firm trading across various asset classes, where some may be complex swaps or debt products, and clients may be investing in portfolios, which puts a very high demand on data collection and a surveillance system’s ability to analyze across the full data set,” said Almqvist of SunGard.

Another requirement is to have rules that adjust its calibration or sensitivity in a way that prevents very high number of false positives to be generated in exceptional market conditions as well as being able to adjust to normal fluctuations of behavior over time, Almqvist said.

Feature image by Cifotart/Dollar Photo Club